@omegaengine/sdk

Decision infrastructure for autonomous AI — TypeScript SDK

The official TypeScript SDK for OmegaEngine. Full type coverage for all 30+ API endpoints including the Security Intelligence Suite.

Installation

npm install @omegaengine/sdk

Quick Start

import { OmegaClient } from "@omegaengine/sdk";

const omega = new OmegaClient({
  baseUrl: "https://api.omegaengine.ai",
  apiKey: process.env.OMEGA_API_KEY!,
});

// Make a decision
const result = await omega.decide({
  scenario: "Should I approve this $50K wire transfer?",
  context: "VIP customer, 780 credit score, 5yr history",
  domain: "finance",
});

console.log(result.ok);                    // true
console.log(result.decision.verdict);      // "ALLOW"
console.log(result.decision.riskLevel);    // "LOW"

Security Intelligence

Access the full CISO-grade security intelligence platform:

// Get the complete security posture (16 engines, ~5ms)
const posture = await omega.security.posture({
  organization: "Acme Corp",
  industry: "fintech",
});

// Adversarial Robustness Score — the FICO score for AI security
const ars = await omega.security.score();
console.log(`ARS: ${ars.report.score}/1000 (${ars.report.bandLabel})`);

// Security SLA tracking
const sla = await omega.security.sla();
if (sla.sla.violated > 0) {
  console.warn(`${sla.sla.violated} SLAs violated!`);
}

// Threat intelligence feed
const threats = await omega.security.threatIntel("critical");
for (const t of threats.catalog) {
  console.log(`[${t.severity}] ${t.name} — ${t.mitre}`);
}

// Zero-day discovery
const zd = await omega.security.zeroDay();

// Anomaly detection
const anomalies = await omega.security.anomalies();

// Chaos engineering
const chaos = await omega.security.chaos();
console.log(`Resilience: ${chaos.chaos.overallResilience}/100`);

CI/CD Security Gate

Block unsafe deploys in your pipeline:

const gate = await omega.security.ciGate({
  minScore: 700,
  failOn: "high",
  sarif: true,
  commitSha: process.env.GITHUB_SHA,
  branch: "main",
});

if (gate.verdict === "FAIL") {
  console.error(gate._ci.annotation);
  process.exit(1);
}

Or use the GitHub Action:

- name: OmegaEngine Security Gate
  uses: omegaengine/omega-gate@v1
  with:
    api-key: ${{ secrets.OMEGA_API_KEY }}
    min-score: 700
    fail-on: high
    sarif: true

Webhooks

Subscribe to real-time security alerts:

// Register a webhook
const wh = await omega.security.registerWebhook({
  url: "https://your-app.com/webhooks/omega",
  events: ["sla.violated", "zero_day.discovered", "anomaly.spike"],
  secret: "whsec_your_signing_secret",
});

// List registered webhooks
const list = await omega.security.listWebhooks();

Available Webhook Events

| Event | Description | |-------|-------------| | sla.violated | Security SLA threshold breached | | sla.at_risk | SLA approaching violation | | anomaly.spike | Anomalous traffic detected | | zero_day.discovered | Novel adversarial pattern found | | attack.campaign_detected | Cross-tenant coordinated attack | | drift.regression | Safety score decreased | | chaos.failure | Chaos experiment failed | | contract.violation | Model behavior breach | | incident.escalated | Incident escalated to P1/P0 | | regulatory.action_required | New compliance requirement |

Full API Coverage

Intelligence

• omega.security.score() — Adversarial Robustness Score
• omega.security.anomalies() — Anomaly detection
• omega.security.forensics() — Model forensics
• omega.security.zeroDay() — Zero-day discovery
• omega.security.threatIntel() — Threat intelligence feed
• omega.security.surface() — Attack surface graph
• omega.security.correlate() — Cross-tenant correlation
• omega.security.fingerprint() — Model fingerprinting
• omega.security.threatModel() — STRIDE threat model

Operations

• omega.security.incidents() — Incident timeline
• omega.security.costs() — Cost attribution
• omega.security.playbooks() — Response playbooks
• omega.security.contracts() — Behavior contracts
• omega.security.sla() — Security SLAs
• omega.security.chaos() — Chaos engineering
• omega.security.simulate() — Threat simulation
• omega.security.tenantRisk() — Tenant risk scoring

Compliance & Reporting

• omega.security.regulatory() — Regulatory tracker
• omega.security.maturity() — Maturity model
• omega.security.sbom() — AI SBOM
• omega.security.pentest() — Pen test report
• omega.security.executiveBrief() — Executive brief
• omega.security.benchmark() — Industry benchmark
• omega.security.drift() — Safety drift detection

Webhooks & CI/CD

• omega.security.registerWebhook() — Register webhook
• omega.security.listWebhooks() — List webhooks
• omega.security.ciGate() — CI/CD security gate

Configuration

const omega = new OmegaClient({
  baseUrl: "https://api.omegaengine.ai",
  apiKey: "omega_your_api_key",
  retry: {
    maxRetries: 3,
    baseDelay: 500,
    maxDelay: 5000,
  },
});

License

Apache-2.0 — © OmegaEngine Inc.